Heat transfer to oil-water flow in horizontal and inclined pipes: Experimental investigation and ANN modeling

Abstract

In this work, local heat transfer coefficients (HTC) and different flow patterns of oil-water two-phase flow in a horizontal and slightly upward inclined (+4° and +7°) pipe were investigated. The test section was an 11 mm inner diameter (ID) copper pipe with a length to diameter ratio of 164. Water and diesel fuel (2.49 mPa s viscosity and 798 kg/m3 density) were selected as immiscible liquids and high speed photography technique was used for the flow pattern identification. The superficial Reynolds numbers ranged from 1350 to 13,700 for water and 300–3700 for oil. The experimental results indicated that the oil-water heat transfer is dependent on the inclination angle and flow pattern. As the pipe inclination angle increases averaged HTC values for each flow pattern increases. It was found that the effect of the flow pattern on the oil-water HTCs is higher than the pipe inclination. In addition, an artificial neural network (ANN) model was developed for predicting the HTC of oil-water two-phase flow in the studied different inclination angles of pipe. Superficial oil Reynolds number (Reso), superficial water Reynolds number (Resw), inclination angles (IA) and some numbers appropriated for each flow pattern (FPN) were selected as input variables, whereas two-phase HTC (hTP) values were selected as output variables. The ANN was trained, validated and tested against the experimental data. The obtained optimal ANN model had good prediction for all of the positions and all flow patterns. Mean absolute percent error (MAPE) of 1.98% and correlation coefficient (R) of 0.993 for testing data set and MAPE of 1.81% and R value of 0.995 for all data sets were achieved.